Aviation Accidents Exploratory Data Analysis¶
In this notebook, we are performing exploratory data analysis on Aviations Accidents Dataset.¶
- Import Standard Libraries
- Load the dataset into a dataframe
- Data Cleanup
- Exploratory Data Analysis
Plot 1: Distribution of Accidents in different Countries (Count of Accidents > 200)
Plot 2: Distribution of Injuries or Fatalities in the World
Plot 3: Distribution of Accidents in the US
Plot 4: Top 10 companies which are involved in accidents
Plot 5: Analysis of Notorious Models of Top 4 Companies involved in accidents
Plot 6: Time Series of Injuries/Fatalities wrt Aircraft.Damage
Plot 7: What was the Phase of Flight and Weather Conditions when accident took place?
Plot 8: What was the Engine Type of Fatal or Non-Fatal accidents? - Observations
- References
In [1]:
#Import Standard Libraries and Packages
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
sns.set(style='darkgrid')
import plotly.express as px
#from shapely.geometry import Point
#import geopandas as gpd
import plotly.graph_objects as go
#from plotly.offline import iplot, init_notebook_mode
#init_notebook_mode()
from warnings import filterwarnings
filterwarnings("ignore")
In [2]:
#Read dataset
aviation1 = pd.read_csv("AviationData.csv", encoding='ISO-8859-1')
In [3]:
#View the dataset
aviation1.head()
Out[3]:
3. Data Cleanup¶
In this notebook, we'll be retaining only the relevant columns for our EDA¶
In [4]:
#Drop irrelevant columns
columns = ['Event.Id', 'Investigation.Type','Accident.Number', 'Airport.Code', 'Airport.Name',
'Registration.Number', 'Air.Carrier', 'Schedule', 'FAR.Description']
aviation2 = aviation1.drop(columns, axis=1)
aviation2.head()
Out[4]:
Column Location contains information about City and State. We'll split this column into two columns. Further, column Injury.Severity contains information about the count of fatalities. We have separate column for that information. So, we'll keep only relevant information in InjurySeverityType field.¶
In [5]:
aviation2['City'] = aviation2['Location'].str.split(',').str[0]
aviation2['State'] = aviation2['Location'].str.split(',').str[1]
aviation2['InjurySeverityType'] = aviation2['Injury.Severity'].str.split('(').str[0]
In [6]:
aviation3 = aviation2.drop(['Location','Injury.Severity'],axis=1)
numeric_columns = aviation3.select_dtypes(include=['float64']).columns
aviation3[numeric_columns] = aviation3[numeric_columns].fillna(0)
aviation = aviation3.fillna("UNKNOWN")
aviation.head()
Out[6]:
Count the number of Accidents as per Countries¶
In [7]:
# Count of Accidents as per Countries
accidents = aviation['Country'].groupby(aviation['Country']).count()
accidents
Out[7]:
In [8]:
# Plot Country
fig = plt.figure(figsize=(20,15))
accidents_Count = accidents[aviation['Country'].groupby(aviation['Country']).count() > 200]
accidents_Count_df = pd.DataFrame({'Country':accidents_Count.index,'Count':accidents_Count.values})
plt.bar(accidents_Count_df['Country'], height = accidents_Count_df['Count'], color='red')
plt.xticks(rotation=90)
plt.xlabel("Country", size=15)
plt.ylabel("Count of Accidents", size=15)
y=accidents_Count_df['Count']
for i,v in enumerate(y):
plt.text(x=i, y=v, s=str(v), horizontalalignment='center', size=15)
plt.title("Distribution of Accidents wrt Countries", size=20)
plt.show()
As depicted from the plot, there are 7 countries namely Australia, Bahamas, Brazil, Canada, Mexico, United Kingdom and the United States which have count of accidents more than 200. Also, most of the accidents have occurred in the US only with staggering 78,849 count.¶
In [9]:
aviation_State = pd.DataFrame(aviation['Total.Fatal.Injuries']+aviation['Total.Serious.Injuries']+aviation['Total.Minor.Injuries'])
aviation_State = aviation_State.rename(columns={0:"Injuries/Fatalities"})
aviation_merged = pd.concat([aviation,aviation_State], axis=1)
aviation_merged.head()
Out[9]:
In [10]:
aviation_merged['text'] = aviation_merged['Country'] + '<br>Injuries/Fatalities ' + (aviation_merged['Injuries/Fatalities'].astype(str))
colors = ["darkblue","yellow","seagreen","purple","cyan","orange"]
limits = [(0,50),(50,100),(100,150),(150,175),(175,200),(350,400)]
#scale = 10
accidentState = []
fig = go.Figure()
for i in range(len(limits)):
lim = limits[i]
Injuries_Count = aviation_merged[((aviation_merged['Injuries/Fatalities'] > lim[0]) & (aviation_merged['Injuries/Fatalities'] <lim[1]))]
fig.add_trace(go.Scattergeo(
locationmode = 'country names',
lon = Injuries_Count['Longitude'],
lat = Injuries_Count['Latitude'],
opacity = 0.8,
text = Injuries_Count['text'],
marker = dict(
size = Injuries_Count['Injuries/Fatalities'],
color = colors[i],
line_color='rgb(40,40,40)',
line_width=0.5,
sizemode = 'area'
),
name = '{0} - {1}'.format(lim[0],lim[1])))
accidentState.append(Injuries_Count)
fig.update_layout(
title_text = 'Distribution of Injuries/Fatalities in the World<br>(Click legend to toggle traces)',
showlegend = True,
geo = dict(
scope = 'world',
landcolor = 'rgb(217, 217, 217)',
)
)
fig.show()
This plot depicts the distribution of Injuries or Fatalities due to aviation accidents in different countries of the world.¶
In [11]:
plt.figure(figsize=(20,15))
aviation_US = aviation[aviation['Country'] == 'United States']
sns.countplot(x='State', data=aviation_US, order = aviation_US['State'].value_counts().index)
plt.xticks(rotation=90)
plt.xlabel("States", size=15)
plt.ylabel("Count of Accidents", size=15)
plt.title("Distribution of Accidents in the United States", size=20)
y=aviation_US['State'].value_counts()
for i, v in enumerate(y):
plt.text(i,v,str(v), horizontalalignment='center', verticalalignment='bottom', fontsize=12, rotation=90)
plt.show()
On diving deep into the distribution of accidents in the different states of US, we have discovered that 'CA' has the highest number of accidents, i.e., 8589 which is followed by 'TX' with 5602 accidents.¶
In [12]:
# Evaluate the Notorious Companies
plt.figure(figsize=(20,15))
aviation['Make'].str.upper().value_counts().sort_values(ascending=False)[:10].plot(kind='bar', color='Green')
plt.xticks(rotation=90)
plt.xlabel("Make", size=15)
plt.ylabel("Count of Accidents", size=15)
plt.title("Top 10 Notorious Companies in the United States", size=20)
y=aviation['Make'].str.upper().value_counts().sort_values(ascending=False)[:10]
for i, v in enumerate(y):
plt.text(i, v, str(v), fontsize=15, style='oblique', horizontalalignment='center')
plt.show()
In this plot, we are analysing the top 10 notorious companies which are involved in most of the accidents. 'Cessna' has 26,073 accidents in the bucket whereas 'Piper' and ' Beech' have 14,230 and 5148 accidents, respectively.¶
In [13]:
# Load the example flights dataset and convert to long-form
data1 = aviation_merged[aviation_merged['Make'].str.upper().isin(["CESSNA","PIPER","BEECH","BELL"])]
data2 = data1.replace('Cessna','CESSNA')
data3 = data2.replace('Bell','BELL')
data4 = data3.replace('Beech','BEECH')
data = data4.replace('Piper','PIPER')
table1 = pd.pivot_table(data, index=['Model'], columns=['Make'], values='Event.Date', aggfunc=np.count_nonzero, fill_value=0)
table2 = table1[(table1.values > 300)]
# Draw a heatmap with the numeric values in each cell
fig, ax = plt.subplots(figsize=(12, 12))
sns.heatmap(table2, annot=True, fmt="d", linewidths=.1, ax=ax, linecolor='white', cmap='YlGnBu')
plt.xlabel("Make", size=15)
plt.ylabel("Model", size=15)
plt.title("Analysis of Notorious Models of Top 4 Companies involved in accidents", size=20)
fig.show()
As shown in the heatmap, CESSNA is the most notorious company. Its model '152' is involved in 2331 accidents. Models '172' and '172N' are involved in 1579 and 1141 accidents, respectively.¶
In [14]:
aviation_merged['Year'] = pd.DatetimeIndex(aviation_merged['Event.Date']).year
aviation_merged['Month'] = pd.DatetimeIndex(aviation_merged['Event.Date']).month_name()
aviation_merged.head()
Out[14]:
In [15]:
fig = px.scatter(aviation_merged[(aviation_merged['Injuries/Fatalities'] > 10) & (aviation_merged['Country'] == 'United States')],
x="Year", y="Injuries/Fatalities", animation_frame="Year", animation_group="State",
size="Injuries/Fatalities", color="Aircraft.Damage", hover_name="State",
size_max=100, range_x=[1980,2020], range_y=[0,400],
category_orders ={'Year':[1962,1979,1980,1981,1982,1983,1984,1985,1986,1987,1988,1989,1990,1991,
1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,
2006,2007,2008,2009,2010,2011,2012,2013,2014,2015,2016,2017,2018,2019,
2020]})
fig.show()
This plot is the animated representation of Injuries or Fatalities with respect to Damages to the aircraft.¶
In [16]:
plt.figure(figsize=(20,15))
splot = sns.countplot(data=aviation_merged, x="Broad.Phase.of.Flight", hue='Weather.Condition', edgecolor=(0,0,0), linewidth=1, palette='Reds')
plt.xticks(rotation=90)
plt.xlabel("Broad Phase of Flight", size=15)
plt.ylabel("Number of Occurrences", size=15)
plt.title("What was the Phase of Flight when Accident took place?", size=20)
for p in splot.patches:
splot.annotate(format(p.get_height()), (p.get_x() + p.get_width() / 2.,
p.get_height()), ha = 'center', va = 'bottom', xytext = (0, 10), textcoords = 'offset points', rotation=90, size=11.5)
plt.show()
This plot illustrates that most of the accidents took place when aircraft was either Landing or Taking off. 19,814 accidents took place when the airplace was Landing and the weather conditions were clear for flight. About 15,385 accidents took place during take off with clear weather conditions.¶
In [17]:
cplot = sns.catplot(x="Engine.Type", col="InjurySeverityType", kind="count", col_wrap=2,
data=aviation_merged[aviation_merged['InjurySeverityType'].isin(['Fatal','Non-Fatal'])], height=4,
legend_out=True, size=8, palette="Set3", edgecolor=(0,0,0))
cplot.set_xticklabels(rotation=90)
cplot.fig.suptitle("What was the Engine Type of Fatal or Non-Fatal accidents?", size=20)
cplot.fig.subplots_adjust(top=.9)
cplot.set_titles("Injury Severity Type is {col_name}")
cplot.set(ylim=[0,65000])
for ax in cplot.axes.ravel():
for p in ax.patches:
ax.annotate(format(p.get_height()), (p.get_x() + p.get_width() / 2,
p.get_height()), ha = 'center', va = 'bottom', xytext = (0, 10), textcoords = 'offset points', rotation=90, size=12)
plt.show()
Post analysis of above two plots, it can said that most of the Fatal or Non-Fatal accident occurred for Engine Type 'Reciprocating'. For Fatal, count is 13,068 whereas for Non-Fatal, count is around 54,310.¶
From the above analysis, we have discovered that:¶
- Most of the aviation accidents have taken place in the United States.
- In the United States, California CA is leading in the count of aviation accidents.
- CESSNA and PIPER are two topmost notorious companies which manufactured aircrafts that are involved in accidents as per this dataset.
- Models 152, 172, and 172N built by CESSNA are most involved in disastrous air accidents.
- Most of the aviation incidents have taken place during Takeoff and Landing. Most of the times, weather conditions were favorable for flight.
- Engine Type of both Fatal and Non-Fatal accidents was Reciprocating for most of the accidents.